The lab

RoboDataLab is an independent research lab — no university behind it, no funding. We work on the deep learning control problems in long-form language: holding state, meaning and character across a whole book rather than a paragraph. We publish what we build.

How we work

We disrupt the future

We are interested in creating a new future, not in optimizing the present. Working on what does not exist yet leaves us free to look for the globally optimal solution instead of a better version of the one already here.

We move slow and with consideration

We spend the time to understand the impact of what we build — the context it will operate in, and what it does to that context — before we commit to it.

We build to last

Nobody would board a vibe-coded airplane or drive over a quick-and-dirty bridge. We treat that kind of shortcut not as time saved but as wasted effort and a fundamental risk.

Who works here, and why

Piotr Trochim

Founder & CEO

Building this lab requires three disciplines converging: deep learning research depth, experience architecting large-scale narrative systems, and over two decades of production software engineering.

Background & Technical Core

  • Deep Learning Research: 5 years at Google DeepMind leading simulation, robotics, and the Control Suite; ads ranking at Meta; founding AI team at Tony Blair Institute. Publications →
  • Narrative Systems: Designed and built the visual quest architecture powering The Witcher 2 and The Witcher 3 at CD Projekt RED, alongside firsthand long-form fiction writing. The Witcher 2: Assassins of Kings →
  • Systems Engineering: 21 years of professional software engineering across C++, Python, JavaScript, TypeScript, C#, Java in telecoms, games, physics, simulation, control, deep learning and large-scale experimentation platforms. Engineering post →

RoboDataLab is focused on building next-generation narrative and language systems powered by deep learning—turning empirical research into software that genuinely works.

Chief Product Officer

Open

The lab provides the technical core. We are looking for an ambitious product counterpart to imagine where this work leads, discover what we should build from first principles, and shape our research into category-defining products.

Who I am looking for

  • You think in paradigm shifts rather than incremental optimizations, imagining how language models can fundamentally reshape creative and computational workflows.
  • You are driven by discovery: you love exploring interaction models, dissecting product design, and translating technical capabilities into intuitive human experiences.
  • You bring high agency and strong product taste as a founding partner—challenging assumptions, prioritizing user utility, and defining what we build from day one.

This is an equal founding seat. If you are hungry to build transformative language, story, and narrative products rather than optimize existing ones, write and introduce yourself.